Presenter: Gábor Szatmári, Department of Soil Mapping and Environmental Informatics
According to the European Journal of Soil Science, the use of mathematical, statistical procedures, tests and analyses in soil research to support the results and to test the hypothesis(es) formulated at the beginning of the research is increasing exponentially. It is known as a rule of thumb in (geo-) statistics that even the best procedures cannot eliminate errors made during sampling. Particular emphasis should therefore be placed on developing a sampling strategy adapted to the issues and hypotheses that are the subject of research.
In our presentation, we will look at concepts inalienably related to sampling strategies that fundamentally determine what conclusions we can make within a given probability model; such concepts include: randomisation, stationarity, sample volume and autocorrelation. In addition, theoretical considerations defining the basis for sampling strategies, such as: probability variable vs. regionalized variable, population vs. superpopulation, model-based vs. plan-based approach are presented. At the end of our presentation, we formulate theoretical questions that could help us answer certain sampling questions.